TRANSPORTATION RESEARCH RECORD llll 71 present standards of uniformity and level of illurninance and 3. M. G. Bassett, S. Dmitrevski, P. C. Kramer, and F. W. Jung. Mea­ luminance in fixed highway lighting cannot reveal a system's surements of Reflection Properties of Highway Pavement Samples. weakness. Journal of the Engineering Society, SepL 1981. 4. J.B. DeBoer and D. A. Schreuder. Theoretical Basis of Road Lighting Design. In Public Lighting, Chap. 3, Philips Technical Library, Eindhoven, Netherlands, 1967. REFERENCES 5. A. Erbay. Atlas of the Reflection Properties of Road Surfaces. Institut fuer Lichttechnik, Technische Universitaet Berlin, Berlin, 1. Determination of Minimum-Maximum Luminance Levels for Pro­ West Germany, 1974. grammable Lighting Operations. Internal Report Ministry of Transportation and Communications, Downsview, Ontario, Can­ ada, 1984. 2. CIB Committee TC-4.6. Calculation and Measurement of Lumi­ nance and Illuminance in Road Lighting. CIB Publication 30 (fC-4.6). Commission Internationale d'Eclairage, Paris, 1976. Publication of this paper sponsored by Committee on V1Sibility. A Method of Calculating the Effective Intensity of Multiple-Flick Flashtube Signals MARC B. MANDLER AND JOHN R. THACKER A method of determining the effective intensity of light flashes A flashtube is a capacitive discharge device capable of emitting composed of multiple pulses (flicks) of light was devised. Detec­ brilliant flashes of light in extremely brief time periods (on the tion thresholds were measured for such flashes when the flick order of microseconds). The highly intense flashtube burst can frequency and flash duration were varied. Thresholds de­ be detected at great distances and has been noted as a conspic­ creased with increasing flick frequency and flash duration. At uous signal in a typical aid-to-navigation system (1). Moreover, each flick frequency the relationship between threshold and flash duration was well characterized by the Blondel-Rey rela­ the efficiency of converting input energy to visible output is tion (a = 0.2), provided a multiplicative frequency-dependent greater than that of an incandescent light (2). These factors fitting parameter was chosen. The fitting parameter, ~. in­ make the flashtube attractive as an aid to navigation. creased linearly with frequency between 5 and 20 Hz. A There are three major disadvantages associated with the use method of determining effective intensity was described that of flashtubes. The intense nature of the flick tends to momen­ uses the flick frequency, number of flicks, and the calculated tarily blind the close observer (3). Also, the duration of the effective Intensity of a single flick to arrive at the solution. It single flick is so brief that mariners have difficulty fixing the was concluded that this method should be used for all multiple­ flick signals, provided the single-flick duration is Jess than 0.01 exact location in the visual field (3 ). Finally, mariners report sec and the frequency is between 5 and 20 Hz. The method of difficulty judging the distance to the flashing source (3 ). The Allard should not be used, because it consistently overesti­ latter two difficulties can be ameliorated by presenting several mates effective intensity. flicks in rapid succession so that the appearance is not one of individual flicks, but of a longer-duration flash. Previous stud­ ies have shown that individuals can take line-of-sight bearings with greater speed and accuracy when the flash duration is increased in this way (4). U.S. Coast Guard Research and Development Center, Avery Point, The detection distance of a lighted aid to navigation is Groton, Conn. 06340-6096. valuable information because it not only allows one to calculate 72 TRANSPOKfATlON RF.sEARCH RECORD llll the range at which a light will become visible, but it is also one DEFINITIONS measure of the signal effectiveness of the aid. A typical aid-to­ navigation system is composed of both steady a.'l.d flashing A single light pulse from a flashtube is called a flick. When lights. The detection distance of a steady light, where intensity several flicks are presented in rapid succession so that dark does not vary with time, can be calculated from the familiar periods are not distinguished between the individual flicks, this Allard's law (5, p. 62): is referred to as a multiple-flick flash. The rate at which the flicks are delivered in a multiple-flick flash is termed the flick frequency. The flash duration or flash length is the time be­ tween onset of the first flick and the cessation of the last flick where and is a function of the flick frequency and the number of flicks. Figure 1 provides two examples of multiple-flick signals. E illuminance threshold of the eye [typically 0.67 sea- The bottom portion of Figure 1 shows relative intensity as a mile candle (6)], function of time for a 20-Hz, 13-flick signal. The time between I intensity of the light, each flick is 1 per flick frequency, which in this case is 0.05 sec. T transmissivity of the atmosphere, and The flash duration is 0.6 sec. ill the upper portion of the figure, D distance at which the light is visible. a 5-Hz, 4-flick signal is shown. Each flick is delivered every 0.2 sec. As with the 20-Ilz signal, the flash duration is 0.6 sec, As the intensity of a flashing light source is a function of time, though the number of flicks and total integrated light intensities the detection distance can be calculated by using Allard's law of the two signals are different. provided that a steady-light-equivalent intensity, termed the effective intensity of the light, can be determined. Effective intensity is defined as follows (7) : 5 Hz If a flash is found to be just seen in conditions in which a steady light of intensity le is also just seen at the same distance and in the same atmospheric conditions, the flash is said to have an effective intensity le. 20 Hz The three generally accepted methods of calculating the effec­ .?• i tive intensity of a single flash are the methods of Allard (5) (not 'ii a: to be confused with Allard's law discussed previously), 0 Schmidt-Clausen (8, 9), and Blondel-Rey-Douglas (10, 11). All 0 .1 .2 .3 .4 .5 .6 .7 .8 .9 these methods are condensed and described by the futemational Time (seconds) Association of Lighthouse Authorities (IALA) (7). FIGURE 1 Plot of intensity versus time of As noted earlier, the increased duration of the multiple-flick multiple-flick flashes. flash is desirable, but to incorporate these multiple-flick flash­ tube devices in an aid-to-navigation system requires a method of specifying its detection range. The IALA formally recog­ nized that of the three single-flash methods of calculating APPARATUS effective intensity, only the Allard method is appropriate for calculating the effective intensity of multiple-flick flashes (12): Figure 2 shows a schematic of the apparatus used for data collection. A flashtube (Automatic Power, Inc., Part The reason for recommending ... the method of Allard for 9001-0295) was installed inside an integrating sphere. A 0.025- trains of rapidly repeated fl ashes was that this method would in. aperture was attached to the output of the integrating sphere yield an effective intensity that increased with increasing num­ and defined the source size. A variable neutral-density wedge ber of flashes in the train and approached asymptotically a steady-state response that for very rapid rates was identical with provided computer control of illuminance of the signal. The Talbot's Law. The other two methods could not yield any signal was reflected off a mirror and superimposed on a low­ satisfactory effective intensity for trains of flashes. It cannot, luminance (0.0045-ft-lambert) background provided by a however, be said that there is any direct experimental confirma­ Kodak slide projector. tion of the effective intensity obtained by the Allard method for The signal appeared as a point source (actual angular diame­ repeated flashes. ter of 0.0003 mrad) when viewed from 83.25 in. The back­ ground subtended approximately 1.0 rad by 1.0 rad. To mini­ The Allard method involves lengthy computer calculations of mize the observer's uncertainty of the position of the signal, the explicit solution of a differential equation (7). The primary four small low-intensity fixation points were provided, each purpose of this work was to find a simple, accurate method of 17.5 mrad from the signal location. determining the effective intensity of multiple-flick signals The fiashtube circuitry was modified so that it could be based on the characteristics of the signal. A secondary goal of triggered by a computer pulse. A single output flick was pre­ this work was to provide the missing experimental confirma­ sented for each input trigger with the maximum rate being 50 tion of the Allard method. flicks per second Mandler and Thack£r 73 Slide projector Variable density! wedge. Integrating sphere ~ "-. Flxation plate / Viewing Aperture with lour surrounding Observer fixation lights FIGURE 2 Schematic of laboratory apparatus. CALIIlRATION Figure 3 shows that the peak output occurs about 20 µsec after flick onset and the intensity decays to 10 percent of peak Figure 3 is a plot of the intensity as a function of time for a after 55 µsec, with negligible low-level output for as long as single near-threshold flick. To obtain this curve, an EG&G about 85 µsec after onset. photomultiplier tube (PMT) with photometric filter was posi­ tioned in the apparatus where the observer's eye was typically positioned. The PMT, which had a rise time of less than IO PROCEDURE nsec, was calibrated against an EG&G Model 555 photometer system using a steady light, so that the relationship between Detection thresholds were obtained for a total of 34 separate illuminance and output voltage from the PMT was known.